Genomics: New Era Begins

Expect more accuracy from estimated breeding values with addition of genomic data to bull proofs

When choosing sires for your herd, you have a new tool at your disposal that greatly enhances estimated breeding value (EBV) accuracy. Genotype information, now included in many EBV calculations, has ushered in a new era for Canadian bull proofs.

Of the bulls listed in the Canadian Dairy Network (CDN) Top 100 Lifetime Profit Index (LPI), for instance, 95 have EBVs that include genotype information. These GEBVs have resulted from a worldwide effort to use basic genetic coding information to improve genetic value estimates of cows and bulls.

The bovine genome is made up of 25,000 to 30,000 genes. In mapping the genome, researchers look for markers called single nucleotide polymorphisms (SNPs). SNPs identify various genetic traits.

How genotyping advanced

Consider SNPs as signs or markers along a road on a genome map. Close to some of these signs are important conditions that affect the outcome of travel down that road. Many signs close together and evenly spaced reveal a lot about the conditions between the signs. In reading the makeup of the bovine genome, many SNPs relatively close together lets researchers predict the outcomes of the gene sequences in that area of the genome.

More than one gene, and sometimes hundreds, control most economically important traits, making gene mapping difficult. After scientists overcame this challenge to achieve the mapping breakthrough, the next big step was developing a computer chip that carries all the coding to recognize most of the identified SNPs. This was the Illumina 50k chip, developed and patented by Illumina Inc.

The first step in applying genomic technology was to genotype a large number of animals. Genotyping can be done on samples of just about any body tissue-blood, semen or hair follicles.

Through the joint project of the U.S. Department of Agriculture, Canadian Dairy Network and many AI studs, a large number of male and female Holsteins were genotyped and used in the initial evaluation.

The first part of actually calculating genomic evaluations was called the learning phase. Genomic evaluations were run with an existing, highlyproven population to determine what the points on the chip actually mean. Since more than one gene controls nearly all economically important traits, a large reference population was needed to accurately measure effects of the SNPs.

One big advantage of the co-operative effort between USDA and Canada to pool all available genotypes for evaluation was accuracy due to the large reference population. More than 5,000 bulls were used to estimate genotype effects.

The final step was taking the genotype information for all new individuals and evaluating them to compare what researchers already knew from the learning phase. The genomic evaluations were then combined with any other available information, such as parent average and progeny proof.

A genomic breeding value can be determined for a calf shortly after birth as a result of genomic testing combined with parent average. It will have a much higher reliability than previously possible.

This allows AI centres to screen young sire entries from a much larger group of bulls than previously possible, providing greater selection and the opportunity to sample a wider variety of bloodlines. The number of bulls actually purchased for progeny testing can be reduced due to the greater accuracy of the selection method.

To harness this knowledge, public sector researchers and major artificial insemination (AI) studs in the USA and Canada launched a massive joint effort. They developed the approach used to produce genomic evaluations, and are among the first in the world to offer this technology to their breeders. CDN published its first evaluations to include genomic data in August.

Sire GEBVs offered for use in your breeding program fall into two categories. The first includes older, progeny-proven sires, and the second covers bulls with no progeny proofs yet.

Blending genomic proofs with their existing progeny data has meant little change for bulls in the first group. If the progeny proof reliability is already over 90 per cent, adding genomic information is of relatively little consequence. There will be some reliability improvement for traits that have lower heritability, such as fertility and calving, and for newly proven bulls with a limited number of firstcrop daughters.

The GEBVs of the second group, with no progeny proofs yet, have reliability ranging from 60 to 65 per cent - much higher than previously possible when only parent average was available. These GEBVs are equivalent to having several daughters in a traditional proof as shown in Table 1. They are still not as accurate as a progeny proof, and most advice is to use them in groups instead of selecting one or two top bulls to use extensively.

Using a group of five bulls in this category has reliability almost as good as that of a proven bull, research has shown. With a group of 10 bulls, the average GEBV has the same reliability as a highly proven bull.

Most of the emphasis in genotyping has been on AI sires and selecting sires for AI use. However, some cows have been genotyped already-mostly tested by AI studs as bull mothers or potential bull mothers. Another relatively small number have been genotoyped as part of a Holstein Canada pilot project.

Table 1: Increased reliability of genomic evaluation over parent average
Trait Parent Average Reliability Genomic Evaluation
Additional Daughter Equivalent
Milk Yield
Fat Yield
Protein Yield
Productive Life
Final Score

If you believe top cows in your herd are candidates for embryo transfer, sale as breeding stock or potentially becoming bull mothers, consider having these animals genotyped. Today's elite cow list includes only those cows that have GEBVs. Those without GEBVs are listed separately.

Genotyping costs can be done through Holstein Canada. See the Holstein Canada website ( for instructions on how to take samples and other procedures and costs.

There is some interest in a smaller genotyping set with much fewer SNPs analysed at a lower cost. This does sacrifice some accuracy. As more testing is done and procedures become more streamlined, however, pricing may become more competitive without reducing accuracy.

GEBVs do not eliminate the need for participating in progeny-proving programs and keeping records of production, calving and other traits. These data are required to update the reference values on which the effects of the various SNPs are estimated. New reference sets will developed and updated on an ongoing basis. Accurate on-farm data form the basis for the program.

Widespread GEBV use could improve genetic progress by as much as 60 per cent. This would be accomplished as AI studs take advantage of more efficient bull proving, producers make better use of younger sires, and genetic evaluation of cows and heifers becomes more accurate.

This appeared in the Ruminations column of The Milk Producer Magazine, October 2009.

For more information:
Toll Free: 1-877-424-1300
Author: Blair Murray - Dairy Genetic Improvement Specialist/OMAFRA
Creation Date: 07 May 2010
Last Reviewed: 07 May 2010